Loading [MathJax]/extensions/MathMenu.js
RL-CNN: Reinforcement Learning-designed Convolutional Neural Network for Urban Traffic Flow Estimation | IEEE Conference Publication | IEEE Xplore

RL-CNN: Reinforcement Learning-designed Convolutional Neural Network for Urban Traffic Flow Estimation


Abstract:

Accurate prediction of urban traffic flows brings enormous advantages to big cities. Therefore, many urban traffic flow predictors have been developed in recent years. Ur...Show More

Abstract:

Accurate prediction of urban traffic flows brings enormous advantages to big cities. Therefore, many urban traffic flow predictors have been developed in recent years. Urban traffic flow predictors aim to identify complex mobility patterns and capture urban traffic flow characteristics from large-scale historical datasets. Afterward, trained models are used to predict the future traffic volume in terms of the number of moving objects (e.g., vehicles). Convolutional Neural Networks (CNN) and other deep learning approaches are brilliant choices because of their ability to learn Spatio-temporal dependencies. Nevertheless, the extensive set of hyper-parameters tends to make these neural networks overly complex and challenging to design. In this work, we introduce RL-CNN, a framework based on Reinforcement Learning whose objective is to autonomously discover highperformance CNN architectures for the given traffic prediction task without human intervention. We examine the proposed RL-CNN model as a traffic flow estimator on a real-world and large-scale vehicular network dataset. We observe improvements of 5% - 10% in the average traffic flow prediction accuracy over the state-of-art approaches.
Date of Conference: 28 June 2021 - 02 July 2021
Date Added to IEEE Xplore: 09 August 2021
ISBN Information:

ISSN Information:

Conference Location: Harbin City, China

I. Introduction

Comprehensive urban traffic information can benefit urban citizens' daily life and improve urban transportation efficiency. Accurate predictions of such traffic information are of great importance for route planning, navigation, and other mobility services. Urban traffic prediction generally applies traffic models to analyze various historical and real-time traffic data to predict traffic conditions in terms of the number of moving objects (e.g., vehicles) in the future. Thanks to the popularity of ubiquitous sensing and Intelligent Transportation Systems (ITS) in recent years, we can gather unprecedented mobility data by exploiting various mobile devices (e.g., smartphones and on-board GPS devices). Such emerging big data availability makes accurate traffic predictions viable.

References

References is not available for this document.